1819.] of the Theory of Heat. 191 



The one which we have chosen appears to us to unite all the 

 requisite conditions. 



It is founded upon the laws of cooling. It is known that 

 there exist between the velocity of cooling of different bodies 

 placed in the same circumstances and the specific heats of the 

 same bodies, relations, in consequence of which the ratio of the 

 capacities may be deduced from that of the times of cooling. 

 The fir^t application of this principle was by Mayer, who satisfied 

 himself that the capacities determined in this .way differ little 

 from those obtained for the same bodies by the method of mix- 

 ture. Mr. Leslie, who has adopted the method of Mayer, has 

 pointed out an additional precaution, of which the latter did not 

 suspect the necessity ; namely, to enclose the body on which 

 we operate in an envelope, which niu^ be always the same, in 

 order to avoid the error which would result from any inequality 

 in the radiating power of the surfaces. But the most important 

 of all the causes of uncertainty, and to which neither Mayer nor 

 Leslie paid any attention, is that which results from the unequal 

 conductibility of the substances compared with each other. The 

 infiuence of this cause is so much the less, the smaller the volume 

 is of the bodies operated upon, and the slower the heat makes 

 its escape from it. Our object then must be to fulfil these two 

 conditions ; but it is difficult to reconcile them, because when 

 we diminish the mass of a body, we augment the velocity with 

 which its heat is dissipated. However, by endeavouring to unite 

 all the causes which contribute to retard the cooling of a given 

 mass, we are enabled, as the experiments have shown, to place 

 it in such circumstances that the difference in the conductibility 

 of the substances operated on has no longer any sensible influence 

 on the measure of the capacities. 



The first method which presents itself for attaining that end is 

 not to begin the observation till the temperature of the body is 

 only a few degrees higher than that of the surrounding bodies. 

 Accordingly all our experiments were made in an interval of 

 temperature included between 10° and 5° centigrade of excess 

 above the ambient medium. It is indispensable to measure the 

 changes of temperature with the greatest possible care ; for even 

 a slight error in the estimation might occasion a great mistake 

 in the result which it is the object to obtain. By operating, as 

 we have said, at the same temperature for all the bodies, we avoid 

 errors resulting from the graduation of the thermometer; and by 

 observing this instrument through a glass, we can increase the 

 size of its degrees so much as not to commit an error exceeding 

 the 50th of a degree, which occasions a degree of uncertainty 

 respecting the specific heat that may be overlooked. It is well 

 known that all these precautions would be delusive if the temper- 

 ature of the ambient medium were not rigorously the same in 

 each case, and during the total duration of every experiment : 



